Applications of Geostationary Data for Operational Forest Fire Monitoring in Brazil Global Geostationary Fire Monitoring Applications Workshop EUMETSAT Darmstadt, Germany March 23-25 Wilfrid Schroeder 1 João Antônio Raposo Pereira 1 Alberto Setzer 2 1 PROARCO/IBAMA 2 CPTEC/INPE [email protected]
19
Embed
Applications of Geostationary Data for Operational Forest Fire Monitoring in Brazil Global Geostationary Fire Monitoring Applications Workshop EUMETSAT.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Applications of Geostationary Data for Operational Forest Fire Monitoring in Brazil
Global Geostationary Fire Monitoring Applications Workshop
Surface Characteristics:(i) Reflectivity (albedo) > 24%(ii) Water: 21x21 matrix having at least one pixel over 80%(iii) Water: 21x21 matrix having at least one pixel over 60% and Tb4 > 15K(iv) Reflective soils: 9x9 matrix having 25% of pixels with Tb2 > 45oC(v) Clouds: 3x3 matrix having 75% of pixels with albedo > 24%
Image Characteristics:(i) Night detection having over 300 hot spots(ii) 50 hot spot night time increase from latest synoptic hour(iii) Over 2000 hot spots during day time images (10:45h-23:45UTC)
Bad lines:(i) Any line having 10+ hot spots over ocean waters(ii) 50 neighbour pixels processed as fire(iii) 300 hot spots along the same line(iv) 97% of Vis Channel pixels having DN=0
CPTEC/INPE Web Product
CPTEC/INPE Web Product
Output Sample File
Nr Lat Lon LatDMS LongDMS Date Time Sat Mun State Country Veg Suscept Prec DWR Risk Persist
1 0.95 -62.7167 N 0 57 0.00 O 62 43 0.00 20040207 84500 GOES-12 Barcelos AM Brasil OmbrofilaDensa BAIXA 24 0 0.1 0
2 1.1 -62.7333 N 1 6 0.00 O 62 43 60.00 20040207 84500 GOES-12 Barcelos AM Brasil OmbrofilaDensa BAIXA 24 0 0.1 0
3 -12.9167 -38.6167 S 12 55 0.00 O 38 37 0.00 20040207 114500 GOES-12 Itaparica BA Brasil OmbrofilaDensa BAIXA 23.6 0 0 0
4 -9.383 -38.2333 S 9 22 60.00 O 38 13 60.00 20040207 114500 GOES-12 Paulo Afonso BA Brasil NaoFloresta MEDIA 0.9 10 0.8 0
5 -8.55 -40.2 S 8 33 0.00 O 40 12 0.00 20040207 114500 GOES-12 Lagoa Grande PE Brasil NaoFloresta MEDIA 0 10 0.9 0
6 -7.983 -40.3167 S 7 58 60.0 O 40 19 0.00 20040207 114500 GOES-12 Ouricuri PE Brasil NaoFloresta MEDIA 0 10 0.9 0
7 -0.016 -62.6167 S 0 1 0.00 O 62 37 0.00 20040207 144500 GOES-12 Barcelos AM Brasil NaoFloresta BAIXA 5 9 0.4 0
8 -0.016 -62.6333 S 0 1 0.00 O 62 37 60.00 20040207 144500 GOES-12 Barcelos AM Brasil Contato BAIXA 27.5 0 0 0
9 0 -62.6333 S 0 0 0.00 O 62 37 60.00 20040207 144500 GOES-12 Barcelos AM Brasil Contato BAIXA 27.5 0 0 0
10 0.05 -62.6167 N 0 3 0.00 O 62 37 0.00 20040207 144500 GOES-12 Barcelos AM Brasil NaoFloresta BAIXA 5 9 0.4 0
Automatic Fire Detection – Case Study
Barcelos
Amazonas
2004
Noaa_12
Noaa_16
MODIS
GOES-12
Total area burned:18000ha
Fire in Barcelos Jan-Feb 2004
0
5
10
15
20
25
30
35
40
2004
0126
2004
0127
2004
0128
2004
0129
2004
0130
2004
0131
2004
0201
2004
0202
2004
0203
2004
0204
2004
0205
2004
0206
2004
0207
2004
0208
2004
0209
2004
0210
Tota
l Mis
sing
Nu
mb
er
of H
ot S
po
ts
Noaa_16
Noaa_12
MODIS
GOES
Automatic Fire Detection – Case Study
Automatic Fire Detection – Continental Scale
Conclusions
• Image usefulness for visual identification of fires is outstanding and proves to be essential to any operational fire monitoring system
• Overall performance of automatic detection is still questionable• Balancing “conservative” x “liberal” algorithms/thresholds
would be desirable – is it attainable?• Field validation should be reinforced and aimed by different
groups – let’s optimize efforts and resources• If we are to consider realistic numbers of active fires being
detected, we must continue (and improve) use of geostationary imagery integrating their fire products to other systems (polar orbiting)